Mixture models; missing data; EM algorithm; 3-way data sets
Dr L. A. (Lyn) Hunt MSc DPhil Waikato
Phone +64 7 838 4038 (Ext. 8338)
Fax +64 7 838 4155
Lyn initially joined the department in 1985 as a graduate student before taking on the role of Tutor in 1988 and Senior Tutor in 1995. Lyn completed a doctorate degree and attained a lecturing position in the Department of Statistics in 1996. Previously Lyn had worked in actuarial positions in the Government Life Office and in an insurance company and also as a chemist monitoring air pollution in New Zealand.
For several years Lyn has been part of the first year teaching team for 'Introduction to Statistical Methods' (0655.121). This
course is mainly taken by students from the schools of Science, and Computing and Mathematical Sciences. The course has an emphasis
on statistical thinking, i.e. interpretation of the computer output for the questions of interest.
In 2002 Lyn is also part of the teaching team for Statistics for Quality Improvement (0655.352A). She will also be teaching Advanced Data Analysis (0655.321B).
Lyn’s research over the past years concerns the clustering of observations using a finite mixture models approach, and where the EM algorithm is employed in the fitting of the model.The model can accommodate both categorical and continuous variables. (Jorgensen and Hunt (1996), Hunt and Jorgensen 1999). Lyn has written a FORTRAN program, Multimix, to estimate the parameters of the distributions when fitting a mixture of K groups, by using this approach.
The approach has been extended so that missing observations can also be accommodated in the model. The program implementing this approach will be released shortly.
Lyn has worked with Associate Professor Kaye Basford, University of Queensland on extending this clustering approach to three-way data sets for mixed datasets (Hunt & Basford 1999). As missing data values are also a problem in three-way data sets, the approach has been extended for data sets with continuous attributes (Hunt & Basford 2001). Our latest work involves extending the technique used for clustering three way data so that these techniques can cope with missing values for datasets with mixed attributes.
Hunt, L.A. & Jorgensen, M.A., (2003). ‘Mixture model clustering for mixed data with missing information’. Computational Statistics & Data Analysis, 41, 429-440.
Hunt, L.A. & Basford, K.E., (2001). ‘Fitting a Mixture Model to Three-mode Three-way Data with Missing Information’. Journal of Classification, 18, 209-226.
Bolstad, W.M., Hunt, L.A., McWhirter, J.L. (2001). ‘Sex, Drugs, and Rock & Roll Survey in a First-Year Service course in Statistics’, American Statistician, 55, No. 2, 145-149.
Hunt, L.A. & Basford, K.E., (1999). ‘Fitting a Mixture Model to Three-mode Three-way Data with Categorical and Continuous Variables’. Journal of Classification, 16, 283-296.
Hunt, L.A. & Jorgensen, M.A., (1999). ‘Mixture Model Clustering: a Brief Introduction to the MULTIMIX Program’. Australian and New Zealand Journal of Statistics, 41, No. 2, 153-171.
McQueen, R.J., Holmes, G. and Hunt, L.A., (1998). ‘User Satisfaction with the Use of Machine Learning as a Data Analysis Method in Agricultural Research’.N.Z. Journal of Agricultural Research, 41, 577-584.
Jorgensen, M.A. & Hunt, L.A. (1996). ‘Mixture Model Clustering of Data Sets with Categorical and Continuous Variables’. - In Proc. of the conference, Information, Statistics and Induction in Science, pages 375 – 384, Eds., D. L. Dowe, K. B. Korb, and J. J. Oliver, World Scientific.
Jorgensen M.A. & Hunt L.A.(1995). Mixture Model Clustering of Data Sets with Categorical and Continuous Variables - Waikato University Research Report No. 36 Series II.
Hunt, L.A. (1994). Mixture Model Clustering of Data Sets with Categorical and Continuous Variables - In Proc. of the Conference, NZSA/ORSNZ.
An activity that I have taken up recently with my family is kayaking. So far this has been mainly done in the sea at the Bay of Islands, but we are now going out at Lake Karapiro.
I also enjoying sea fishing. I do have to admit that I generally don’t catch much, and my largest catch has been well and truly outsized by my daughter Kelly’s catch on 1/1/2002.
I enjoy ‘pottering away’ in our garden in my spare time. I am trying to get a
‘maintenance free’ garden (if such a thing is possible).